In Peskin and Schroeder's book p 83, the authors said

At any fixed time $t_0$, we can of course expand $\phi$ in terms of ladder operators: $$\phi(t_0, \textbf{x})=\displaystyle\int\frac{d^3p}{(2\pi)^3}\frac{1}{\sqrt{2E_{\mathbf{p}} }}(a_{\mathbf{p}} e^{i\textbf{p.x}}+a_{\mathbf{p}} ^\dagger e^{-i\textbf{p.x}}) (1) $$ ... $e^{iH_0 (t-t_0)} \phi (t_0, \mathbf{x}) e^{-iH_0 (t-t_0)} \equiv \phi_I(t,\mathbf{x}).$...

Since we can diagonize $H_0$, it is easy to construct $\phi_I$ explicitly:

$\phi_I(t, \mathbf{x} ) = \int \frac{d^3p}{ (2\pi)^3} \frac{1}{ \sqrt{2 E_{\mathbf{p}}} } \left( a_{\mathbf{p}} e^{-ipx} + a^{\dagger}_{\mathbf{p}} e^{ipx} \right) |_{x^0 = t - t_0} $ (4.15)

To derive (4.15), I thought about $$ e^{i H_0 (t-t_0)} a_{\mathbf{p}} e^{-i H_0 (t-t_0)} $$ use BCH expansion in terms of $e^{-B} A e^B$ to work out (4.15), similar to the free field theories.

As answered in Interacting Fields in QFT $a_{\mathbf{p}}$ is time-indenpendent for free theory (I thought about $(\partial^2 + m^2) \phi(x) =0 $ in Eq. (2) in the aforementioned link for free theories) but time-dependent on interacting theories. Will this difference affect commutation relation $[a^{int}_{\mathbf{p}},H_0]$? More specifically, $$ [a^{int}_{\mathbf{p}},H_0 = \int \frac{d^3p }{(2\pi)^3} \omega_p a_{\mathbf{p}}^{free, \dagger} a_{\mathbf{p}}^{free} ] =? $$


1 Answer 1


By considering the Heisenberg-picture, we can deduce that given a free Hamiltonian $H_0$ and the annihilation operator $a_{\bf{k}}$, we have that $$[a_{{\bf{k}}},H_0]=\omega_{{\bf{k}}}a_{{\bf{k}}}\implies a_{\bf{k}}(t)=e^{iH_0t}a_{{\bf{k}}}e^{-iH_0t}=e^{-i\omega_{\bf{k}}t}a_{\bf{k}}.$$Here $a_{\bf{k}}(t)$ is the time dependent annihilation operator. Since $\omega_{\bf{k}}\in\mathbb{R}$, we have that $e^{-i\omega_{\bf{k}}t}\in\mathbb{C}$. We can use this to compute the commutation relation between $a_{\bf{k}}(t)$ and $H_0$. We thus have$$[a_{\bf{k}}(t),H_0]=[e^{-i\omega_{\bf{k}}t}a_{\bf{k}},H_0]=e^{-i\omega_{\bf{k}}t}[a_{\bf{k}},H_0]=\omega_{\bf{k}}e^{-i\omega_{\bf{k}}t}a_{\bf{k}}.$$But we have already established that $e^{-i\omega_{\bf{k}}t}a_{\bf{k}}=a_{\bf{k}}(t)$. Thus we have that$$[a_{\bf{k}}(t),H_0]=\omega_{\bf{k}}a_{\bf{k}}(t).$$ So we can see that the time-dependent annihilation (for example) commutates with $H_0$ in the same manner as the time-independent annihilation operator. Intuitively, this makes sense; if the Hamiltonian doesn't change with time, then at all times, the creation and annihilation operators will create/annihilate the same eigenstates $|k\rangle$ with corresponding energy eigenvalues $\omega_{\bf{k}}$.

It should be noted that this is not true in general. For a general Hamiltonian $H$, the commutation relations might be far more complicated. Furthermore, equal-time commutation relations can be shown to preserve the unitary structure of the operators and unequal-time commutation relations may vary.

  • $\begingroup$ Thanks. My concern is, the difference betwen $a^{int}_p$ and $a^{free}_p$, based on Eq. (2) in the answer in physics.stackexchange.com/questions/310353/…, is the non-zero part of $(\partial^2 + m^2)\phi(x) = H_{int} \phi(x)$. The term triggerd by $H_{int}$ can possibly be complicated. I am not sure if we can have $a_k(t) = e^{- i \omega_k t} a_k$ for $a_k^{int}$ $\endgroup$
    – AlphaF20
    Dec 28, 2021 at 20:59
  • $\begingroup$ $a_{{\bf{k}}}e^{-iH_0t}=e^{-i\omega_{\bf{k}}t}a_{\bf{k}}$ is derived from free field, I think. $\endgroup$
    – AlphaF20
    Dec 28, 2021 at 21:16
  • $\begingroup$ You can show that $e^{iH_0t}a_{{\bf{k}}}e^{-iH_0t}=e^{-i\omega_{{\bf{k}}}t}a_{{\bf{k}}}$. In the answer you referenced, it is shown that the time derivative of $a_{{\bf{k}}}$ is non-zero only when the Hamiltonian has an interacting component. The $a_{\bf{k}}$ (free) is constant in time when your theory is described only by a free field $H_0$. Whilst $a_{\bf{k}}(t)$ will in general depend on the interacting Hamiltonian and will not necessarily be the same for all time. $\endgroup$ Dec 29, 2021 at 3:00
  • $\begingroup$ That's my concern. I can derive $e^{iH_0t}a_{{\bf{k}}}e^{-iH_0t}=e^{-i\omega_{\bf{k}}t}a_{\bf{k}}$, in a way similar to P&S (2.46). By using $e^-B A e^B = A + [A, B] + \cdots$, noticing $[A,B] = [a_{\mathbf{k}}, -iHt] $, and P&S (2.32) $[H, a_{\mathbf{k}} ] = -\omega_{\mathbf{k}} a_{\mathbf{k}}$. But, that's all for free fields. My question is about if the above derivation is also valid in interacting theories, for specifically, $[a^{free}_{\mathbf{k}}, a^{int, \dagger}_{\mathbf{k}'}] = ?$. I'm not sure if I overlooked your point. $\endgroup$
    – AlphaF20
    Dec 29, 2021 at 4:53
  • $\begingroup$ Right, I see the confusion. Formally, the $e^{iH_0t}a_{\bf{k}}e^{-iH_0t}$ is called the interaction picture operator because when the coupling constant in $H_{\text{int}}$ is small (and usually it is assumed to be), most of the time dependence in $a_{\bf{k}}(t)$ will come from $H_0$, hence my previous arguments. If you want to describe the entire field, you then impose unitary operators that have $H$ and $H_0$ both involved. At that point I believe that closed form commutation relations seize to exist. $\endgroup$ Dec 29, 2021 at 6:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.